NASA comments on agile vs. waterfall comparing SpaceX and Boeing [pdf] | Hacker News Reader

NASA comments on agile vs. waterfall comparing SpaceX and Boeing [pdf]

oiir.hq.nasa.gov

From the pdf:

As the ASAP has pointed out previously, SpaceX and Boeing have very different philosophies in terms of how they develop hardware. SpaceX focuses on rapidly iterating through a build-test-learn approach that drives modifications toward design maturity. Boeing utilizes a well-established sys- tems engineering methodology targeted at an initial investment in engineering studies and analysis to mature the system design prior to building and testing the hardware. Each approach has advantages and disadvantages.

Basically this is a big win for "Agile" (which is about as losely defined as possible now)

This is a significant condemnation of the "systems design" approach that has been central to space engineering so far. It might have been appropriate in the past, when access to space was actually expensive, but it has also significantly stifled the progress within established players, and only a new player, free from orthodox thinking patterns, was able to innovate.

Personally I think it is more like "old player that did not understand software development gets trounced by new player that has software at its core"

This is not waterfall vs agile really. But software eats the world

But the remarkable thing is, the agile software methodology is equally applicable to non-software. SpaceX does the same with engines & rockets. Tesla does the same with cars. That's the real breakthrough.

Yes. But. To my mind that's the 'iterate' or 'fail fast' methodology - which is pretty much de facto in Science and Engineering as a best practise. See benchtests for motors and rockets, or firing chickens at jet engines. Of course it is not always achieved - but is that because scientists and engineers do not know about it, or because the organisation in which they work is hobbling them in some fashion?

I think there are two big points here - one, software development is a core skill for firms these days, without which you are likely to fail.

Second, "heavy engineering" appears to have hit a point where the leading edge is not progressing as fast as the "commercial" world is. (this might be part of the feeling we are not inventing the future fast enough anymore).

By this I mean once upon a time, the envelope was so far out only States could afford to keep up - Canada folded its aerospace efforts in the 50s etc. As a rough rule of thumb you needed to not only build custom components, but you needed to build custom machines to make the custom components.

Today I think commercially available machines are able to make the custom components, thus making the envelope much more within reach for much less money - its that Makers can now do things previously only professionals could, so now professionals can do things previously only big organisations could.

This means that the inefficiencies of big 'dinosaur' organisations no longer pay off (Boeing is not the only people who can do rockets).

Elon Musk seems to have known / discovered / stumbled (#) this inflection point - but it is likely to play out in all sorts of industries. Military equipment etc. I mean if a 300 USD drone is a viable weapon on urban asymmetric battlefields, then something is coming (##)

(#) select based on personal views of Musk

(##) my personal publically obvious inflection point is a "maker style" rocket that can damage a jet aircraft. The rocket community is heading towards rockets that can get to the 32,000 foot level, and once you do that at Mach 1, for a few thousand dollars in a backyard, air dominance becomes something up for grabs again.

Reply on news.ycombinator.com